CN103123691B - A kind of filter method and device of Moire fringe - Google Patents
A kind of filter method and device of Moire fringe Download PDFInfo
- Publication number
- CN103123691B CN103123691B CN201310060300.4A CN201310060300A CN103123691B CN 103123691 B CN103123691 B CN 103123691B CN 201310060300 A CN201310060300 A CN 201310060300A CN 103123691 B CN103123691 B CN 103123691B
- Authority
- CN
- China
- Prior art keywords
- moire fringe
- image
- detection
- preset
- filtering
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000001514 detection method Methods 0.000 claims abstract description 70
- 238000001914 filtration Methods 0.000 claims abstract description 50
- 238000009499 grossing Methods 0.000 claims abstract description 29
- 230000008569 process Effects 0.000 claims abstract description 10
- 238000003708 edge detection Methods 0.000 claims description 14
- 238000004458 analytical method Methods 0.000 claims description 9
- 230000003044 adaptive effect Effects 0.000 claims description 6
- 230000003362 replicative effect Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 230000000694 effects Effects 0.000 abstract description 7
- 208000012260 Accidental injury Diseases 0.000 abstract description 3
- 208000014674 injury Diseases 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 5
- 238000002474 experimental method Methods 0.000 description 4
- 238000010191 image analysis Methods 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 230000035755 proliferation Effects 0.000 description 1
- 230000000452 restraining effect Effects 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The present invention provides a kind of filter method of Moire fringe and devices, and wherein method includes: S1, thens follow the steps S2 if detecting Moire fringe to current image progress Moire fringe detection；If Moire fringe is not detected, terminate process；S2, it after carrying out Gaussian smoothing to current image using preset window, goes to step S1 and carries out the filtering of next round Moire fringe, until reaching preset wheel number；The preset window size wherein used in step s 2 in the filtering of each round Moire fringe is incremented by.The accidental injury of when without Moire fringe image is reduced to image to present invention degree as big as possible, also be reduced as far as selecting because of filter type it is improper caused by filtering is incomplete or filtered the problem of working as, improve the image entirety recognition effect based on Moire fringe filtering.
Description
[technical field]
The present invention relates to computer application technology, in particular to the filter method and device of a kind of Moire fringe.
[background technique]
With the fast development of mobile network's technology and the rapid proliferation of intelligent terminal, user more and more passes through intelligence
Terminal surfs the Internet to obtain information, and optical character identification (ORC) technology is wherein to help user that fast acquiring information is facilitated to commonly use
One of mode.For example, will be shone for being taken pictures using the camera function of intelligent terminal to character content by wireless network
Carry out automatic identification after piece upload server, then recognition result returned into intelligent terminal by wireless network, with facilitate user into
The further service such as row search.
Currently used application scenarios are that user takes pictures paper document, the label etc. that content is outdoor scene, but there is also one kind
Typical application scenarios are that user is answered using the character on intelligent terminal shooting computer screen or mobile phone screen etc. for this
Then there is larger difficulty with the OCR of scene.Because from screen shot to image will receive the interference of Moire fringe, Moire fringe
Presence drastically influence the identification accuracy of OCR, therefore it is very necessary and crucial for carrying out Moire fringe filtering before OCR
's.
The methods of image airspace smoothed analysis, image frequency domain analysis and image similarity analysis is primarily present at present to come to
Know that the image comprising Moire fringe eliminates the interference of Moire fringe, but existing these types of mode on the one hand can be to no Moire fringe
Normal picture cause biggish accidental injury, when on the other hand carrying out disposable filtering to the Moire fringe of image, if filtering side
Improper be likely to result in of formula selection filters incomplete or filtered the problem of working as, to influence whole recognition effect.
[summary of the invention]
In view of this, the present invention provides a kind of method and apparatus for filtering Moire fringe, in order to reduce to no More
The accidental injury of the normal picture of striped improves whole recognition effect.
Specific technical solution is as follows:
A kind of filter method of Moire fringe, this method comprises:
S1, S2 is thened follow the steps if detecting Moire fringe to current image progress Moire fringe detection；If
Moire fringe is not detected, then terminates process；
S2, it after carrying out Gaussian smoothing to current image using preset window, goes to step S1 and carries out next round More item
The filtering of line, until reaching preset wheel number；
The preset window size wherein used in step s 2 in the filtering of each round Moire fringe is incremented by.
A preferred embodiment according to the present invention, it is described that Moire fringe detection is carried out in the following ways to current image
At least one of:
Mode one: replicating current image, carries out Gauss using the preset window of epicycle to the image of duplication and puts down
It is sliding, obtain smoothed out image；It counts pixel value in the error image of current image and smoothed out image and is greater than default picture
The ratio R1 of plain value Td；If R1 is greater than or equal to preset ratio threshold value T1, no to current image detection to Moire fringe
Moire fringe is then not detected；
Mode two carries out edge detection to current image, obtains edge image；Connected domain point is carried out to edge image
Analysis counts the number N of thin-and-long connected domain, and wherein thin-and-long connected domain is that the length-width ratio of circumscribed rectangle is higher than default length-width ratio threshold
The connected domain of value；If N is greater than or equal to preset number threshold value T2, it is determined that the current image detection to More's item
Otherwise Moire fringe is not detected in line.
A preferred embodiment according to the present invention, if the preset wheel number is 2, in the mistake of first round Moire fringe
Employing mode one carries out Moire fringe detection in filter, and employing mode two carries out Moire fringe in the filtering of the second wheel Moire fringe
Detection.
A preferred embodiment according to the present invention, the Td are 20, T1 0.15, and T2 is relevant to edge image length and width
Adaptive number.
A preferred embodiment according to the present invention, if the preset wheel number is 2, in the mistake of first round Moire fringe
Use length for 5 in filter, the preset window that variance is 0.8 uses length for 7, variance in the filtering of the second wheel Moire fringe
For 1.7 preset window.
A kind of filter device of Moire fringe, the device include:
Detection unit, for carrying out Moire fringe detection to current image；
Smooth unit, for when the detection unit detects Moire fringe, using preset window to current image
Carry out Gaussian smoothing；
Control unit, it is single for being supplied to the detection using the smooth unit treated image as present image
Member, until reaching preset wheel number；When Moire fringe is not detected in the detection unit, terminate the processing of the device；
The preset window size that wherein smooth unit described in the filtering of each round Moire fringe uses is incremented by.
A preferred embodiment according to the present invention, the detection unit is at least one of in the following ways to current figure
As carrying out Moire fringe detection:
Mode one: replicating current image, carries out Gauss using the preset window of epicycle to the image of duplication and puts down
It is sliding, obtain smoothed out image；It counts pixel value in the error image of current image and smoothed out image and is greater than default picture
The ratio R1 of plain value Td；If R1 is greater than or equal to preset ratio threshold value T1, no to current image detection to Moire fringe
Moire fringe is then not detected；
Mode two carries out edge detection to current image, obtains edge image；Connected domain point is carried out to edge image
Analysis counts the number N of thin-and-long connected domain, and wherein thin-and-long connected domain is that the length-width ratio of circumscribed rectangle is higher than default length-width ratio threshold
The connected domain of value；If N is greater than or equal to preset number threshold value T2, it is determined that the current image detection to More's item
Otherwise Moire fringe is not detected in line.
A preferred embodiment according to the present invention, if the preset wheel number is 2, in the mistake of first round Moire fringe
Detection unit employing mode described in filter one carries out Moire fringe detection, detects described in the filtering of the second wheel Moire fringe single
First employing mode two carries out Moire fringe detection.
A preferred embodiment according to the present invention, the Td are 20, T1 0.15, and T2 is relevant to edge image length and width
Adaptive number.
A preferred embodiment according to the present invention, if the preset wheel number is 2, in the mistake of first round Moire fringe
Smooth unit described in filter uses length for 5, the preset window that variance is 0.8, described in the filtering of the second wheel Moire fringe
Smooth unit uses length for 7, the preset window that variance is 1.7.
As can be seen from the above technical solutions, the present invention carries out Moire fringe detection and Gaussian smoothing by wheel to realize to figure
As reducing to image to degree as big as possible the mistake of image when without Moire fringe from thin Moire fringe to the filtering of thick Moire fringe
Wound, be also reduced as far as selecting because of filter type it is improper caused by filtering is incomplete or filtered the problem of working as, mention
The high image entirety recognition effect based on Moire fringe filtering.
[Detailed description of the invention]
Fig. 1 is the broad flow diagram of the filter method of Moire fringe provided in an embodiment of the present invention；
Fig. 2 is a kind of method flow diagram for detecting Moire fringe provided in an embodiment of the present invention；
Fig. 3 is the method flow diagram of another detection Moire fringe provided in an embodiment of the present invention；
Fig. 4 is a kind of preferred embodiment flow chart of Moire fringe filter method provided in an embodiment of the present invention；
Fig. 5 is the broad flow diagram of the filter device of Moire fringe provided in an embodiment of the present invention；
Fig. 6 a is an original image example provided in an embodiment of the present invention；
Fig. 6 b is the result that the image provided in an embodiment of the present invention to Fig. 6 a is directly identified；
Fig. 6 c is that the image provided in an embodiment of the present invention to Fig. 6 a carries out the filtered result of Moire fringe；
Fig. 6 d is the result provided in an embodiment of the present invention identified on the basis of Fig. 6 c；
Fig. 7 a is another original image example provided in an embodiment of the present invention；
Fig. 7 b is the result that the image provided in an embodiment of the present invention to Fig. 7 a is directly identified；
Fig. 7 c is that the image provided in an embodiment of the present invention to Fig. 7 a carries out the filtered result of Moire fringe；
Fig. 7 d is the result provided in an embodiment of the present invention identified on the basis of Fig. 7 c.
[specific embodiment]
To make the objectives, technical solutions, and advantages of the present invention clearer, right in the following with reference to the drawings and specific embodiments
The present invention is described in detail.
It assumes that image to be processed is the image comprising Moire fringe in view of existing Moire fringe filter method, treats
The image of processing is directly filtered Moire fringe, on the one hand this mode can cause the normal picture for not including Moire fringe
It accidentally injures, when on the other hand carrying out disposable filtering to the Moire fringe of image, is likely to result in if filter type selection is improper
Filter incomplete or filtered the problem of working as.Therefore Moire fringe detection is carried out to image first in the present invention, only for
Image comprising Moire fringe carries out the filtering of Moire fringe.And in the detection of Moire fringe by the way of classification, i.e.,
Using taking turns from carefully filtering to thick Moire fringe, can all the detection of Moire fringe be carried out in each round, if detected not more
You are then filtered in epicycle striped, if Moire fringe is not detected, due to being from thin Moire fringe gradually to thick More
Bar detection, once thinner Moire fringe is not detected, it may be considered that current image no longer includes Moire fringe, it is subsequent
Thicker Moire fringe detection it is also no longer necessary, terminate the process of filtering.
Main implementation process is as shown in Figure 1, comprising the following steps:
Step 101: using original image as current image, starting to execute subsequent step.
Step 102: Moire fringe detection being carried out to current image and is thened follow the steps if detecting Moire fringe
103；If Moire fringe is not detected, terminate process.
Current image refers to original image, the i.e. content of step 101 at the beginning, since the second wheel, currently
Image be exactly image after last round of Gaussian smoothing.When carrying out Moire fringe detection, it can use but be not limited to following
Two ways:
First way is as shown in Fig. 2, mainly comprise the steps that
Step 201: current image being replicated, Gauss is carried out using the preset window of epicycle to the image of duplication and is put down
It is sliding, obtain smoothed out image.
It is actually pre- Gaussian smoothing to be carried out to present image, and pass through subsequent step 202 and step 203 in this step
Judge the difference of the image before the image and Gaussian smoothing after Gaussian smoothing, determines that present image is according to the difference
No includes Moire fringe.Therefore the preset window used here is identical as the preset window that step 103 in epicycle uses.
Step 202: counting pixel value in the error image of current image and smoothed out image and be greater than presetted pixel value
The ratio R1 of Td.
Value I in error image at coordinate (x, y)diff-If(x, y) are as follows:
Idiff-If(x,y)=abs(I(x,y)-Isf(x, y)), wherein I (x, y) is in current image at coordinate (x, y)
Value, Isf(x, y) is the value in smoothed out image at coordinate (x, y), and abs () expression takes absolute value.
Step 203: if R1 is greater than or equal to preset ratio threshold value T1, to current image detection to Moire fringe,
Otherwise Moire fringe is not detected to current image.
In this approach, presetted pixel value Td and proportion threshold value T1 can be chosen for empirical value or experiment value.
The second way is as shown in figure 3, mainly comprise the steps that
Step 301: edge detection being carried out to current image, obtains edge image.
The edge detection used in this step can be but be not limited to Canny edge detection, and Canny edge detection is
A kind of more common edge detection method, generally comprise 4 steps: 1) gaussian filtering is to image denoising；2) by original gradation figure
Find out 2 gradient maps in length and breadth, and comprehensive gradient map；3) non-maximum restraining is carried out in conjunction with 3 gradient maps；4) edge company is carried out
It connects.It has been existing more mature technology in view of Canny edge detection, details are not described herein.Side has just been obtained after detecting edge
The image that edge is surrounded, i.e. edge image.
Step 302: connected domain analysis being carried out to edge image, counts the number N of thin-and-long connected domain, wherein thin-and-long connects
Logical domain is the connected domain that the length-width ratio of circumscribed rectangle is higher than default length-width ratio threshold value.
Step 303: if N is greater than or equal to preset number threshold value T2, it is determined that arrived to the current image detection
Otherwise Moire fringe is not detected in Moire fringe.
Regard Moire fringe as thin-and-long connected domain in this mode to detect, if the number of thin-and-long connected domain compared with
It is more, reach preset number threshold value T2, then it is assumed that current image detection to Moire fringe, because normal picture is to include
Greater number of thin-and-long connected domain.Number threshold value T2 can use empirical value or experiment value.
The analysis of connected domain generally comprises two steps: 1) connected component labeling, i.e., gives under the definition of defined connectivity
Each object or Target Assignment one unique non-zero label；2) shape feature of each connected domain, institute of the present invention are calculated
The shape feature of finger is circumscribed rectangular characteristic；The number of shape feature statistics thin-and-long connected domain according to each connected domain.
Prior art is used in view of connected domain analysis, details are not described herein.
It should be noted that the mode of above two Moire fringe detection can use phase in the filtering of each wheel Moire fringe
With mode, can also each wheel Moire fringe filtering in different ways, such as in the first round use the first not
The mode of your bar detection, in the second wheel in such a way that second of Moire fringe detects.
Step 103: after carrying out Gaussian smoothing to current image using preset window, going to step 102 and carry out next round
The filtering of Moire fringe, until reaching preset wheel number.Wherein in the filtering of the Moire fringe of each round, the default window of use
Mouth size is incremented by, i.e. the gradually detection and Gaussian smoothing from the detection and Gaussian smoothing of thin Moire fringe to thick Moire fringe.
Once detecting Moire fringe, Gaussian smoothing is used to the mode that Moire fringe is filtered.Implement in the present invention
In example since the filtering of thin Moire fringe, i.e., Gaussian smoothing is first carried out using wicket, then again in next round Gauss filtering,
If still detecting the presence of Moire fringe, then increase window and carry out Gaussian smoothing, slightly thicker Moire fringe is filtered out；Such as
Moire fringe is not detected in fruit, because thin Moire fringe does not all detect, then it is considered that Moire fringe is not present.It is this
Mode similar to " throwing a stone to clear the road " is the damage in order to reduce Gaussian smoothing to the greatest extent to image, is first gradually increased since wicket
It is big to carry out Gaussian smoothing, until filtering out the Moire fringe in image.For there is only the images of thin Moire fringe, then only with
The Gaussian smoothing of wicket can filter out Moire fringe, so as to avoid carrying out Gaussian smoothing using excessive big window
Image bring is damaged.
A kind of preferred embodiment is described by taking the filtering of two-wheeled Moire fringe as an example below, thin More is carried out in the first round
The detection of striped and Gaussian smoothing carry out detection and the Gaussian smoothing of thick Moire fringe, process such as Fig. 4 institute in the second wheel
Show, mainly comprises the steps that
Step 401: obtaining original image I.
Step 402: after replicating to original image I, Gauss being carried out using preset wicket to the image of duplication
Smoothly, smoothed out image I is obtainedsf。
Preset wicket can use such as length for 5 herein, the window that variance is 0.8.
Step 403: determining original image I and smoothed out image IsfError image Idiff-If。
Step 404: statistics difference image Idiff-IfMiddle pixel value is greater than the ratio R1 of presetted pixel value Td, in the value of this Td
Such as 20 can be taken.
Step 405: if ratio R1 is greater than or equal to T1, the value of T1 can take such as 0.15, then detect original figure
As I contains thin Moire fringe, execution step 406；Otherwise terminate process.
Step 406: original image I is replaced with into smoothed out image Isf, complete the filtering of thin Moire fringe.
Step 407: to smoothed out image IsfCanny edge detection is carried out, edge image E is obtainedsf。
Step 408: to edge image EsfConnected domain analysis is carried out, the number N of long and narrow connected domain is counted.
Step 409: if N is greater than or equal to preset number threshold value T2, can be set to and edge image in this T2
The relevant adaptive number of length and width, it is determined that image IsfComprising thick Moire fringe, step 410 is executed；Otherwise terminate process.
Step 410: to image IsfGaussian smoothing is carried out using preset big window, which can use all
If length is 7, the window that variance is 1.7 obtains smoothed out image IIf, so as to complete the filtering of thick Moire fringe.
It is the detailed description carried out to method provided by the invention above, the embodiment of the present invention is provided below with reference to Fig. 5
Device be described in detail.As shown in figure 5, the filter device of Moire fringe may include: detection unit 01, smooth unit 02
With control unit 03.
Detection unit 01 is used to carry out Moire fringe detection to current image.When initial, current image is original
Image, in the Moire fringe filtering of subsequent rounds, current image is last round of treated image.
Specifically, detection unit carries out Moire fringe detection to present image at least one of in the following ways:
Mode one: replicating current image, carries out Gauss using the preset window of epicycle to the image of duplication and puts down
It is sliding, obtain smoothed out image；It counts pixel value in the error image of current image and smoothed out image and is greater than default picture
The ratio R1 of plain value Td；If R1 is greater than or equal to preset ratio threshold value T1, no to current image detection to Moire fringe
Moire fringe is then not detected.
Value I in error image at coordinate (x, y)diff-If(x, y) are as follows: Idiff-If(x,y)=abs(I(x,y)-Isf(x,
Y)), wherein I (x, y) is the value in current image at coordinate (x, y), Isf(x, y) is coordinate (x, y) in smoothed out image
The value at place, abs () expression take absolute value.In this approach, presetted pixel value Td and proportion threshold value T1 can be chosen for empirical value
Or experiment value.
Mode two carries out edge detection to current image, obtains edge image；Connected domain point is carried out to edge image
Analysis counts the number N of thin-and-long connected domain, and wherein thin-and-long connected domain is that the length-width ratio of circumscribed rectangle is higher than default length-width ratio threshold
The connected domain of value；If N is greater than or equal to preset number threshold value T2, it is determined that current image detection to Moire fringe,
Otherwise Moire fringe is not detected.
Wherein carrying out edge detection can be but be not limited to Canny edge detection, and above-mentioned number threshold value T2 can be adopted
With empirical value or experiment value.
Smooth unit 02 be used for when detection unit 01 detects Moire fringe, using preset window to current image into
Row Gaussian smoothing.
Control unit 03 is used to be supplied to detection unit 01 using treated the image of smooth unit 02 as present image, directly
To reaching preset wheel number；When Moire fringe is not detected in detection unit 01, terminate the processing of the device, exports current figure
As can be used for the application such as subsequent image recognition as treated image.It is put down in the filtering of above-mentioned each round Moire fringe
The preset window size that sliding unit 02 uses is incremented by.
Have the mode of being preferably carried out, process shown in corresponding diagram 4, preset wheel number is 2, then in first round Moire fringe
Filtering in 01 employing mode one of detection unit carry out Moire fringe detection, second wheel Moire fringe filtering in detection unit
Employing mode two carries out Moire fringe detection.Each threshold value chosen: Td 20, T1 0.15, T2 are and edge image length and width phase
The adaptive number closed.Smooth unit 02 uses length for 5 in the filtering of first round Moire fringe, the default window that variance is 0.8
Mouthful, smooth unit 02 use length for 7 in the filtering of the second wheel Moire fringe, variance for 1.7 preset window.
After the filtering of above-mentioned Moire fringe, it will be largely increased to the whole recognition effect of image, in Fig. 6 a
Shown in for original image, if directly identified to the original image, recognition result may be to show shown in Fig. 6 b
Right recognition effect is very poor.If after being detected using the method and apparatus provided in the embodiment of the present invention to the original image,
It detects it includes thin Moire fringe and after carrying out Gaussian smoothing, the filtered image of Moire fringe i.e. Fig. 6 c is obtained, in the figure
The recognition result carried out on the basis of image shown in 6c is as shown in fig 6d, it is clear that recognition effect is greatly improved.
Again shown in Fig. 7 a for original image, if directly identified to the original image, recognition result
It may be that can not identify correct result completely shown in Fig. 7 b.If using the method and dress that are provided in the embodiment of the present invention
It sets after being detected to the original image, detect in the first round it includes thin Moire fringe and carries out Gaussian smoothing, second
Wheel is detected it includes thick Moire fringe and after carrying out Gaussian smoothing, the filtered image of Moire fringe i.e. Fig. 7 c is obtained, at this
The recognition result carried out on the basis of image shown in Fig. 7 c is as shown in figure 7d, it is clear that recognition result is greatly improved.
In several embodiments provided by the present invention, it should be understood that disclosed device and method can pass through it
Its mode is realized.For example, the apparatus embodiments described above are merely exemplary, for example, the division of the unit, only
Only a kind of logical function partition, there may be another division manner in actual implementation.In addition, in each embodiment of the present invention
Each functional unit can integrate in one processing unit, be also possible to each unit and physically exist alone, can also be two
Or more than two units are integrated in one unit.
The above-mentioned integrated unit being realized in the form of SFU software functional unit can store and computer-readable deposit at one
In storage media.Above-mentioned SFU software functional unit is stored in a storage medium, including some instructions are used so that a computer
It is each that equipment (can be personal computer, server or the network equipment etc.) or processor (processor) execute the present invention
The part steps of embodiment the method.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (Read-
Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic or disk etc. it is various
It can store the medium of program code.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Within mind and principle, any modification, equivalent substitution, improvement and etc. done be should be included within the scope of the present invention.
Claims (8)
1. a kind of filter method of Moire fringe, which is characterized in that this method comprises:
S1, S2 is thened follow the steps if detecting Moire fringe to current image progress Moire fringe detection；If do not examined
Moire fringe is measured, then terminates process；
S2, it after carrying out Gaussian smoothing to current image using preset window, goes to step S1 and carries out next round Moire fringe
Filtering, until reaching preset wheel number；
The preset window size wherein used in step s 2 in the filtering of each round Moire fringe is incremented by,
It is described that Moire fringe detection is carried out at least one of in the following ways to current image:
Mode one: replicating current image, carries out Gaussian smoothing using the preset window of epicycle to the image of duplication, obtains
Obtain smoothed out image；It counts pixel value in the error image of current image and smoothed out image and is greater than presetted pixel value Td
Ratio R1；It is not otherwise examined if R1 is greater than or equal to preset ratio threshold value T1 from current image detection to Moire fringe
Measure Moire fringe；
Mode two carries out edge detection to current image, obtains edge image；Connected domain analysis, system are carried out to edge image
The number N of thin-and-long connected domain is counted, wherein thin-and-long connected domain is that the length-width ratio of circumscribed rectangle is higher than default length-width ratio threshold value
Connected domain；If N is greater than or equal to preset number threshold value T2, it is determined that from the current image detection to Moire fringe,
Otherwise Moire fringe is not detected.
2. the method according to claim 1, wherein if the preset wheel number is 2, in first round More
Employing mode one carries out Moire fringe detection in the filtering of striped, and employing mode two carries out in the filtering of the second wheel Moire fringe
Moire fringe detection.
3. T2 is long with edge image the method according to claim 1, wherein the Td is 20, T1 0.15
The relevant adaptive number of width.
4. the method according to claim 1, wherein if the preset wheel number is 2, in first round More
Use length for 5 in the filtering of striped, the preset window that variance is 0.8 uses length in the filtering of the second wheel Moire fringe
It is 7, the preset window that variance is 1.7.
5. a kind of filter device of Moire fringe, which is characterized in that the device includes:
Detection unit, for carrying out Moire fringe detection to current image；
Smooth unit, for being carried out to current image using preset window when the detection unit detects Moire fringe
Gaussian smoothing；
Control unit, for being supplied to the detection unit using the smooth unit treated image as present image, directly
To reaching preset wheel number；When Moire fringe is not detected in the detection unit, terminate the processing of the device；
The preset window size that wherein smooth unit described in the filtering of each round Moire fringe uses is incremented by,
The detection unit carries out Moire fringe detection to present image at least one of in the following ways:
Mode one: replicating current image, carries out Gaussian smoothing using the preset window of epicycle to the image of duplication, obtains
Obtain smoothed out image；It counts pixel value in the error image of current image and smoothed out image and is greater than presetted pixel value Td
Ratio R1；It is not otherwise examined if R1 is greater than or equal to preset ratio threshold value T1 from current image detection to Moire fringe
Measure Moire fringe；
Mode two carries out edge detection to current image, obtains edge image；Connected domain analysis, system are carried out to edge image
The number N of thin-and-long connected domain is counted, wherein thin-and-long connected domain is that the length-width ratio of circumscribed rectangle is higher than default length-width ratio threshold value
Connected domain；If N is greater than or equal to preset number threshold value T2, it is determined that from the current image detection to Moire fringe,
Otherwise Moire fringe is not detected.
6. device according to claim 5, which is characterized in that if the preset wheel number is 2, in first round More
Detection unit employing mode one described in the filtering of striped carries out Moire fringe detection, the institute in the filtering of the second wheel Moire fringe
It states detection unit employing mode two and carries out Moire fringe detection.
7. device according to claim 5, which is characterized in that the Td is 20, T1 0.15, and T2 is long with edge image
The relevant adaptive number of width.
8. device according to claim 5, which is characterized in that if the preset wheel number is 2, in first round More
Smooth unit described in the filtering of striped uses length for 5, the preset window that variance is 0.8, in the mistake of the second wheel Moire fringe
Smooth unit described in filter uses length for 7, the preset window that variance is 1.7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310060300.4A CN103123691B (en) | 2013-02-26 | 2013-02-26 | A kind of filter method and device of Moire fringe |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310060300.4A CN103123691B (en) | 2013-02-26 | 2013-02-26 | A kind of filter method and device of Moire fringe |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103123691A CN103123691A (en) | 2013-05-29 |
| CN103123691B true CN103123691B (en) | 2019-02-12 |
Family
ID=48454664
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310060300.4A Active CN103123691B (en) | 2013-02-26 | 2013-02-26 | A kind of filter method and device of Moire fringe |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103123691B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106131391B (en) * | 2016-08-31 | 2019-03-22 | 宇龙计算机通信科技(深圳)有限公司 | A kind of image pickup method and device |
| CN108921823B (en) * | 2018-06-08 | 2020-12-01 | Oppo广东移动通信有限公司 | Image processing method, apparatus, computer-readable storage medium and electronic device |
| CN109889696B (en) * | 2019-03-18 | 2021-06-25 | 上海顺久电子科技有限公司 | Anti-noise shot image recognition method and system for automatic geometric correction |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101263417A (en) * | 2005-09-30 | 2008-09-10 | 夏普株式会社 | Liquid crystal display device and television receiver |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI258023B (en) * | 2001-11-07 | 2006-07-11 | Ibm | A prism sheet, a back-light unit using said prism sheet, and a transmission type liquid crystal display device |
-
2013
- 2013-02-26 CN CN201310060300.4A patent/CN103123691B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101263417A (en) * | 2005-09-30 | 2008-09-10 | 夏普株式会社 | Liquid crystal display device and television receiver |
Non-Patent Citations (1)
| Title |
|---|
| 数码相片莫尔条纹消除算法的研究;余佳;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20070615(第6期);第34-54页 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103123691A (en) | 2013-05-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109325954B (en) | Image segmentation method and device and electronic equipment | |
| US9549121B2 (en) | Image acquiring method and electronic device | |
| CN110233971B (en) | A shooting method and terminal, and a computer-readable storage medium | |
| Dirik et al. | Source camera identification based on sensor dust characteristics | |
| CN109446942A (en) | Method for tracking target, device and system | |
| CN108769731A (en) | The method, apparatus and electronic equipment of target video segment in a kind of detection video | |
| CN103870824B (en) | A kind of face method for catching and device during Face datection tracking | |
| CN107256407B (en) | A kind of hyperspectral remote sensing image classification method and device | |
| CN109584198B (en) | Method and device for evaluating quality of face image and computer readable storage medium | |
| CN111047556B (en) | Strip steel surface defect detection method and device | |
| CN103310439B (en) | The most steady extremal region detection method of a kind of image based on metric space | |
| CN104061749A (en) | Refrigerator food storage position recording method, device and terminal and refrigerator | |
| CN106258009A (en) | A kind of gather the method for fingerprint, fingerprint capturer and terminal | |
| CN103123691B (en) | A kind of filter method and device of Moire fringe | |
| WO2017177559A1 (en) | Image management method and apparatus | |
| CN107209556B (en) | System and method for processing depth images capturing interaction of an object relative to an interaction plane | |
| CN108765532A (en) | Children paint this method for establishing model, reading machine people and storage device | |
| CN111986155A (en) | Universal scene character recognition method and device | |
| CN111669492A (en) | Method for processing shot digital image by terminal and terminal | |
| WO2024212998A1 (en) | Video person re-identification method and system and related device | |
| CN109977875A (en) | Gesture identification method and equipment based on deep learning | |
| CN115062186B (en) | Video content retrieval method, device, equipment and storage medium | |
| US9141854B2 (en) | Method and apparatus for generating structure of table included in image | |
| CN110889817A (en) | Image fusion quality evaluation method and device | |
| CN109359689A (en) | A kind of data identification method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |